Grantee Research Project Results
Final Report: Defining and Predicting PCB Fluxes and Their Ecological Effects in Stream and River Systems for Risk Characterizations
EPA Grant Number: R832213Title: Defining and Predicting PCB Fluxes and Their Ecological Effects in Stream and River Systems for Risk Characterizations
Investigators: Burton, Jr., G. Allen , Ren, Jianhong-Jennifer
Institution: Wright State University - Main Campus , Texas A & M University - Kingsville
EPA Project Officer: Hahn, Intaek
Project Period: March 1, 2005 through February 29, 2008 (Extended to February 28, 2009)
Project Amount: $325,000
RFA: Greater Research Opportunities: Persistent, Bioaccumulative Chemicals (2004) RFA Text | Recipients Lists
Research Category: Hazardous Waste/Remediation , Land and Waste Management , Safer Chemicals
Summary/Accomplishments (Outputs/Outcomes):
Overview: To accurately determine ecological risk and effective remedial actions, it is necessary to understand how ecosystem dynamics affect the linkage of exposure and ecological effects. Without this understanding, risk cannot be determined with certainty. In particular, a fundamental process that must be quantitatively understood is the flux of contaminants from sediments, a major exposure compartment for benthic organisms, into overlying water and biota. This research demonstrated improved characterizations and the prediction of how solids-associated PCB (and related compound) exposures affect aquatic organisms. This investigation of PCB-type fluxes between sediments and overlying water characterized the dominant flux process of advection, through laboratory experiments and theoretical modeling. Laboratory experiments were conducted in stream recirculating flume systems (SRF) using a range of sediment types (low to high levels of gravel, sand, clay, and organic matter) and flow conditions. Chemicals in the SRF were loaded from above to simulate suspended solids loadings of PCB and from below to simulate groundwater upwellings and resuspension events. Experiments were conducted with dichlorodiphenyldichloroethylene (DDE), as a PCB surrogate, for health safety reasons. The SRF was used to calibrate the model and allow for simultaneous characterizations of bioaccumulation and adverse biological effects. Test organisms were exposed to pore waters, surficial sediments, overlying waters and associated colloids, and suspended sediments within the recirculating systems. Bioaccumulation was measured with solid phase microextraction (SPME) (replacing Tenax), and with Lumbriculus. The scope of this project did not allow for a complete field validation of the model, however.
The SRF at Texas A&M University at Kingsville (TAMUK) was used to conduct a flume run with polydispersed fine sediments. Analytical methods for determination of dichlorodiphenyldichloroethylene (DDE) and poly-chlorinated biphenyl (PCB) concentrations were developed and have been tested and the use of SPME evaluated. The amount of DDE sorbed on the SPME fiber depends on the sorption time. Once the system of the fiber and water reaches equilibrium, an increase in sorption time does not increase the amount of DDE sorbed on the fiber. To determine the sorption time required to reach equilibrium, fiber sorption kinetics was evaluated for a DDE solution of 70 µg/L. No significant difference in the peak area was observed among the sorption times tested. Thus, 30 minutes of sorption provides sufficient time for the system to reach equilibrium. Since significant carryover was not observed during the fiber cleaning, 2 minutes of desorption time was sufficient to prevent carryover of DDE and incomplete desorption of DDE from the SPME fiber. Sorption kinetics of PCB Aroclor 1242 were evaluated using batch adsorption tests. PCB adsorption was evaluated in batch sorption experiments. Various techniques were evaluated to determine adsorption rates using sand and montmorillonite clay with various static to mixing techniques. In general equilibrium was reached with 3 to 4 days, which was similar to field studies using in situ exposures of Lumbriculus in PCB contaminated sediments (Burton et al., 2005).
The first case for modeling was the downwelling case, following the transport of PCBs from the water column into the sediment bed. To predict the amount of PCBs adsorbed on the bed sediments and in the pore water, the multiphase transport model developed by Ren and Packman (2004) was modified to output the contaminant distribution in the streambed.
Results of the downwelling study on exchange of DDE between streams and streambeds indicated that the presence of kaolinite or natural colloidal particles significantly influenced the exchange of p,p’-DDE. This exchange process was modeled using a two-site kinetic/equilibrium sorption/desorption model in addition to the physical exchange process. The sorption/desorption model input parameters were estimated using independent batch sorption/desorption experiments.
Results of upwelling of p,p′-DDE between streams and streambeds for base flow condition indicated that the effect of water chemical conditions on DDE release is closely related to the scale of the experiments conducted and the composition and concentration of ionic strength. It slightly increased as pH decreased in the batch experiments conducted; however experimental error may have masked this effect in the flume experiments. Increases in NaCl concentrations generally decreased p,p′-DDE release in batch and flume experiments; while increases in CaCl2 in the batch experiments did not have a significant effect on p,p′-DDE release. A greater release of p,p′-DDE was observed in the presence of NaCl than CaCl2 for batch experiments. A fast release of DDE was observed in the 1st hour under all experimental conditions which was unexpected. A fast release was not observed for Sb release from contaminated sediments collected from a tributary of the Rio Grande, so it was hypothesized that the DDE contaminated sediments used in this project have a significant fraction of loosely bound DDE, due to their being spiked (Northcott and Jones 2000ab and USEPA 2001). Since desorption is the only mechanism used to describe the release of DDE from sediments in the model, the observation of the fast DDE release at the beginning of the experiments makes the model application and verification more complicated.
The upwelling of DDE can be modeled using the modified advective pumping theory and a two-stage desorption model. By incorporating a bioconcentration factor (BCF), the uptake of DDE by benthic organisms at certain locations was effectively simulated and predicted.
Unfortunately, repeated attempts to collect field contaminated PCB sediments were unsuccessful at the Hudson River, Dicks Creek, Fox River, Grasse River and Kalamazoo River. Though these sites have been documented as having PCB contaminated sediments they could not be obtained from surficial grabs with ponar dredge samplers. If additional experiments could be conducted in the future (both in the laboratory and field) with PCB contaminated sediments, the above hypothesis and models could be better tested.
The upwelling and downwelling models developed via this project represent major sediment flux processes and provide a promising tool to predict the behavior of PBTs in natural streams and thereby allow for more accurate predictions of exposure and thus risk. However, to date, the current models have only been verified using experimental data collected using DDE.
Student Support
Several post-doctoral, doctoral and MS students participated in this project at TAMUK and Wright State University. At WSU this included Dr. Keith Taulbee, Preston Smith, and Xueling Zhang. Due to a lack of adequate USEPA funding for their support, the personnel were jointly supported by other discretionary funds of Dr. Burton’s. At TAMUK, students included Celina Camarena, Sivacharan Peddireddy and Doris Otero. Celina was been supported by the NSF funded Center for Research Excellence in Science and Technology (CREST)-Research on Environmental Sustainability of Semi-Arid Coastal Areas (RESSACA). Sivacharan was supported by this STAR grant from June 2005 to April 2007. Doris started to work on this project in September 2007 and has been partially supported by this STAR grant. In addition, one visiting research scholar, Dr. Tong Zheng, worked on the modeling of resuspension of contaminated fine sediments from May 2008 to Sept. 2008.
Results
Downwelling case
The downwelling case study was conducted by investigating the transport behavior of p,p’-DDE between streamwater and streambed in the presence of naturally occurring fine sediments such as kaolinite clay particles. The experiments were conducted using a recirculating flume. The solid-phase microextraction (SPME) method, combined with gas chromatography, was used to determine the concentrations of DDE. The first flume run was a control run which consisted of only DI water and p,p’-DDE to examine p,p’-DDE mass loss due to the flume materials. The second flume run performed was to examine the stream-subsurface exchange of p,p’-DDE. This flume run consisted of DI water, cleaned Ottawa #12 Flint silica sand (F12) (bed sediment), and p,p’-DDE. The third, fourth, and fifth flume runs were conducted to investigate the effects of fine sediments on p,p’-DDE contaminant transport. These flume runs consisted of DI water, cleaned F12 silica sand (bed sediment), p,p’-DDE, and kaolinite particles or natural fine sediments. The kaolinite particles used had an average diameter of 0.22 µm, which was measured using a Brookhaven Instrument Corporation (BIC) ZetaPALS Multi-Angle particle size analyzer. The natural fine sediments were collected from the Rio Grande River.
The multiphase exchange model for the downwelling case was developed by considering the kinetic adsorption/desorption of DDE to bed sediments and suspended sediments, particle deposition and hydrodynamic exchange. The two-site equilibrium/kinetic model proposed by Van Genuchten and Wagenet (1989) and verified using the batch experiments conducted in this study was applied to model the kinetic interaction of DDE with sediments. The results are summarized in a series of figures presented in Appendix 1 and discussed below.
Figure 1 shows the batch experiment results for the interaction of DDE with silica sand. A fast sorption of p,p’-DDE to silica sand within the first 3-5 days followed by a slow equilibrating process was observed. The DDE sorption to silica sand was modeled using the kinetic sorption/desorption model with fitting parameters of f = 0.92 and a = 0.34 day-1, where f is the fraction of equilibrium sites and a is the mass transfer rate.
Figure 2 shows the isotherm data collected from batch experiments after 54 days. The sorption behavior of p,p’-DDE was described using the Freundlich Isotherm (nonlinear case) as compared to the Linear Isotherm with fitting parameters of kd = 0.071 L/g and ns = 1.88, where kd is the empirical distribution coefficient and ns is nonlinear Freundlich coefficient.
Sorption kinetics of PCB Aroclor 1242 were evaluated using batch adsorption tests. Various techniques were evaluated to determine adsorption rates using sand and montmorillonite clay with various static-to-mixing techniques. In general, equilibrium was reached within 3 to 4 days. These findings match those obtained in field studies where concentrations of PCBs in the tissues of the sediment dwelling oligochaete worm, Lumbriculus variegates, reach equilibrium within 3 to 4 days of being exposed in situ in Dick’s Creek (Burton et al., 2005).
The in situ caged exposure chambers that have been routinely used by Burton et al. (2005) were modified to be much smaller for the SRF exposures, so as to reduce their disruption of flow and related artifacts. This also allowed for greater spatial resolution of exposure variation from surficial sediments to overlying waters. Exposures of multiple organisms within the SRF experiments showed no toxicity and little bioaccumulation due to the low levels of PCBs found in the field collected sediments.
Figure 3 shows the comparison of flume results obtained from flume runs #1 and #2 and model simulations for flume run #2. Results are presented as the normalized in-stream concentrations (C* = C/C0) of dissolved p,p’-DDE plotted as a function of dimensionless time. Results from flume run #1 indicate a significant loss of DDE mass apparently due to the flume channel, which needs to be considered in model applications. Results from flume run #2 show that the net bed exchange of p,p’-DDE with clean F12 silica sand in the absence of suspended fine sediments is insignificant. The nonlinear Freundlich coefficient, ns, obtained from the batch experiments (Figure 2) did not accurately predict the DDE exchange results as expected due to the DDE mass loss in the channel although a fitted ns = 1.22 better represented Flume Run #2 results..
Figure 4 shows the comparison of the dissolved p,p’-DDE exchange results obtained from flume runs #2-4. The significant effect of kaolinite colloidal particles on DDE exchange was clearly demonstrated. These results indicate that fine sediments can significantly affect the stream-subsurface exchange of p,p’-DDE and that additional contaminant immobilization in the streambed can occur due to the deposition of fine solids in sediments.
Figure 5 shows the effect of input parameters on simulating the exchange of DDE between stream and streambed under the experimental conditions of flume Run #2. Figure 5A illustrates that increases in f, the fraction of equilibrium sites, shows no significant difference in the results of p,p’-DDE retention in the sediment. Figure 5B show that increases in kd, the distribution coefficient, result in an increase of p,p’-DDE retention in the sediment. Figure 5C demonstrates that increases in ns, the nonlinear Freundlich coefficient, result in a decrease of p,p’-DDE retention in the sediment. Thus, under the given flume experiment conditions, the ns parameter was the most sensitive parameter.
Overall, results indicate a kinetic DDE sorption/desorption behavior characterized by a fast sorption/desorption within the first 3-5 days followed by a slow equilibrating process. This kinetic behavior can be modeled using a two site kinetic/equilibrium sorption/desorption model. The sorption/desorption model input parameters can be estimated using independent batch sorption/desorption experiments. A significant DDE mass loss occurred due to the flume channel even though preventative measures such as using a PTFE liner were employed. This mass loss should be considered in order to correctly apply the exchange model to simulate the exchange of DDE with the streambed. The introduction of fine particles (kaolinite colloids) significantly influences the exchange of p,p’-DDE between the stream and streambed. Under given flume experiment conditions, the Freundlich nonlinear coefficient, ns, is the most sensitive input parameter.
PCB exchange.
Two downwelling case flume experiments (Flume Run #1 and Flume Run #2) were conducted using apparently PCB contaminated sediments. These experiments were designed to investigate the scenario of PCB contaminated sediments being transported downstream to uncontaminated areas along with corresponding biological effects. In both experiments, Ottawa F12 silica sand was used as the test sediment while PCB associated fine sediment was obtained from the Grasse River, NY. In Flume Run #1, two-centimeter SPME polydimethylsiloxane (PDMS) fiber pieces were used to evaluate bioavailability while in Flume Run #2 five EPA test organisms were utilized. The five surrogate test species that were used included: the midge, Chironomus tentans (now called C. dilutus) (8-12 d post hatch), amphipod, Hyalella azteca (14-21 d), cladoceran, Daphnia magna (48 hr), fathead minnow, Pimephales promelas (14-21 d), and mixed-age oligochaetes, Lumbriculus variegatus. Culturing procedures followed USEPA methods (USEPA 1989, 2000). Each experiment lasted approximately 7 days. All PCB analysis was done by a certified analytical laboratory (Alloway, Lima Ohio).
It was expected initially that the sediment obtained from the Grasse River, NY contaminated with PCBs ranging from 5-10 mg/kg. However, the average concentration of PCB-1260 in the sediment was 0.21 mg/kg. Thus, at this initial concentration there was only 3 µg/L of PCB actually injected into the flume as PCB associated suspended sediment which was well below detection limitsof 5 µg/L for aqueous phase samples. Therefore, there was no PCB detected in neither the stream water samples nor the SPME fibers. However, results for dissolved organic carbon (DOC) and total suspended solid (TSS) concentrations over time were obtained.
Figure 6 shows the results for downwelling Flume Run #1 DOC concentration over time. The initial DOC concentration was 4.9 mg /L and then decreased to 4 mg/L and remained fairly stable over time. The decrease in DOC could have been due to the introduction of suspended sediment where some of the DOC may have attached to the suspended sediment. Figure 7 illustrates the TSS results over time. Results show that the TSS concentration decreased over time suggesting fine sediment depositing over time.
For downwelling Flume Run #2, the same sediment was utilized, thus PCBs were not detected (or below detection limit) in the stream water samples, in the streambed sediment core sample, and in the L.variegatus tissue analysis. Results obtained for Flume Run #2 were for DOC, TSS, turbidity, and survival rates for four test organisms (D. magna, C.tentans, P.pomelas and H. azteca).
Figure 8 shows the results for downwelling Flume Run #2 DOC concentration over time. The initial DOC concentration was 4.9 mg /L and then decreased to 4 mg/L and remained fairly stable over time. The decrease in DOC could have been due to the introduction of suspended sediment where some of the DOC may have attached to the suspended sediment. Figure 9 illustrates the TSS results over time. The TSS concentration values also decreased over time suggesting fine sediments depositing over time. Figure 10 depicts the results for the organism survival in the in-situ chambers in the flume. The survival rates were as follows: D. magna 32%, C.tentans 76%, P.pomelas 95%, and H. azteca 97%. These results suggest that the H. azetca and P.pomelas had the highest survival rate under the high fine sediment concentrations in the flume. Control organism survival results were greater than 80% making the test and the use of organisms valid for this experiment as seen in Table 1.
Table 1. Control Organism Survival Rates
Replicate
|
P. pomelas
|
D. magna
|
C.tentans
|
H. azteca
|
1
|
10
|
19
|
10
|
20
|
2
|
10
|
19
|
10
|
20
|
3
|
10
|
20
|
9
|
20
|
Mean Surv.
|
100.00%
|
96.67%
|
87.88%
|
100.00%
|
Upwelling case
Experimental study for base flow condition. Artificially prepared p,p′-DDE contaminated natural river sediment collected from the Rio Grande located in Laredo, TX, was used as contaminated bed sediment. A field-wet sediment spiking procedure (Northcott and Jones 2000ab and USEPA 2001) was followed. The effects of pH and ionic strength (NaCl and CaCl2) on the release of p,p′-DDE from contaminated streambeds were investigated using both batch and flume experiments. For each batch experiment, 30 grams of spiked p,p′-DDE sediment was mixed in 100 mL 10-800 mM NaCl or CaCl2 solution of varying pHs (pH 5-10) in 500 mL amber glass jars and shaken for 24 hr period. Water and sediment samples were then taken after 24 hrs. For flume experiments, spiked p,p′-DDE sediment was placed into a tilting recirculating flume lined with PTFE at a bed depth of 5-6 cm. Deionized water at the desired water chemistry conditions was then slowly introduced into the flume reaching a total height of 15 cm. Flume pump was turned on to desired stream velocity (16 cm/s) and water samples were taken over time. NaOH and/or HCl was added as needed to help maintain the pH of the water. Parameters were measured over time include pH, conductivity, dissolved p,p’-DDE, p,p’-DDE associated with sediment, DDE-associated with colloids, dissolved organic carbon (DOC), colloid zeta potential, suspended particle size, and colloid concentration. p,p′-DDE was analyzed using SPME (100 μm PDMS fiber) extraction combined with a Shimadzu Gas Chromatographer-2014 (GC-ECD). DOC was measured using a Shimadzu TOC 5000-A analyzer. Colloid concentration was determined by filtering 50 mL water sample with 0.02 μm filter, drying solids at 105 ºC for one hour in oven and weighing the sample afterwards. Zeta potential and suspended particle size were determined using ZetaPALS with Zeta Plus System (Brookhaven Instruments Corp.)
Figure 11 shows representative batch results. More p,p′-DDE is released at lower pHs than at higher pHs. DOC release is enhanced at lower and higher pHs (5.23, 6.12, and 10.23). Effect of ionic strength on p,p′-DDE release is more pronounced in NaCl solution than in CaCl2 solution. p,p′-DDE release decreases as NaCl concentration increases to 500 mM; however, change with increasing CaCl2 is minimal. Similar DOC releases were observed in NaCl and CaCl2 solutions and DOC release increases as ionic strength increases with both NaCl and CaCl2.
Figure 12 shows representative flume results. Increased p,p′-DDE release was observed within the first hour followed by a slow release over time for all flume runs. DOC data for pH 8.52 was not available due to instrument failure. DOC concentrations increased over time for flume experiments conducted at pH 4.97 and pH 7.62. At pH 7.62 and 8.52, 100-250 mg/L colloids were released and the colloids had an average particle size of 0.92 µm and 0.63 µm, respectively. At pH 4.97, 60-100 mg/L colloids were released and the colloids had an average particle size of 0.61 µm. High colloid concentrations contributed to the increased colloidal phase p,p′-DDE at pHs 7.62 and 8.52.
Figure 13 shows that colloid release increased at the higher pHs. DOC release slightly increased with increasing pH. Unlike what was observed in the batch experiments, no significant difference in p,p′-DDE release over the tested pH range was observed in the flume experiments. This absence of pH effect indicates that the slight effect of pH observed in the batch experiments is not distinguishable in the flume experiments.
Figure 14 shows that colloid release decreased with increasing NaCl concentrations. DOC data for 10 mM NaCl was not available due to instrument failure. Increases in NaCl concentrations cause more p,p′-DDE to be retained in the streambed which is consistent with the batch experiment results.
Overall, these results indicate that p,p′-DDE release was observed to slightly increase as pH decreased in the batch experiments; however experimental error may have masked these results in the flume experiments. Increases in NaCl concentrations generally decreased p,p′-DDE release in batch and flume experiments; while increases in CaCl2 in the batch experiments did not have a significant effect on p,p′-DDE release. A greater release of p,p′-DDE was observed in the presence of NaCl than CaCl2 for batch experiments.
Model simulations for base flow condition. A process-based multi-phase reactive transport model simulating the upwelling of Persistent Bioaccumulative Toxic (PBT) organic pollutants from contaminated streambed sediments under low-flow conditions and the biouptake of PBT organic pollutants by benthic organisms was developed. The physical exchange process occurring between the streams and streambeds was modeled using a modified advective pumping theory. The interaction of PBT organic pollutants with bed sediments was modeled using a kinetic two-stage desorption equation. A bioconcentration factor (BCF) equation was used to model the biouptake of PBT organic pollutants by benthic organisms.
The residence time function simulations were compared with those obtained using the multiphase linear sorption/desorption model developed by Ren and Packman (2004). Simulations were conducted using different retardation factors, R. The equivalence of the two-stage kinetic desorption to the linear sorption/desorption at different R values was established by setting desorption rate very fast so that the kinetic effect is minimal. The resulting equivalent parameter values between the two types of sorption/desorption equations presented in Table 2 were used in these simulations.
Table 2. Equivalence of the two-stage desorption and linear sorption/desorption parameters.
R
|
krap (1/hr)
|
kslow (1/hr)
|
fslow
|
frap
|
5
|
10
|
5E-06
|
0.8
|
0.2
|
2.4
|
8
|
5E-08
|
0.6
|
0.4
|
1.4
|
7
|
5E-07
|
0.3
|
0.7
|
1
|
10
|
5E-08
|
0
|
1
|
Figure 15 shows that the residence time function calculated using the new model match well with the ones obtained from the previous model, indicating that no numerical errors occur in the residence time function calculation. The fractions of mass released from the streambed simulated using the modified model of Ren and Packman match well with the ones obtained from the new model.
Figure 16 shows that increase in the fraction of rapid desorption site or decrease in the fraction of slow desorption site results in an increase in the fraction of mass released from the streambed. A high rapid kinetic desorption rate indicates a faster release of the contaminants from the streambed than a low kinetic desorption rate.
Table 3 shows the biouptake results. The Bioconcentration factors BCF for Daphnia were obtained from Geyer et al (1991). Since most of the contaminant is available in dissolved phase, the highest concentration of contaminant is found in the organisms, Corg (case I). For case II, less Corg is noticed since sorption/desorption parameters are applied. In case III, the lowest Corg is obtained since less contaminant is available in the pore water for the organisms. For cases IV and V, high BCF indicates high Corg.
Table 3. Biouptake results.
Case #
|
BCF, L/mg
|
frap
|
fslow
|
Krap, 1/hr
|
Kslow, 1/hr
|
Corg , mg/g
|
I
|
1.32E-02
|
0
|
0
|
0
|
0
|
4.43E-07
|
II
|
1.32E-02
|
0.84
|
0.16
|
0.69
|
0.14
|
2.71E-07
|
III
|
1.32E-02
|
0.02
|
0.98
|
0.03
|
3.74E-04
|
1.46E-08
|
IV
|
1.85E-02
|
0.02
|
0.98
|
0.03
|
3.74E-04
|
2.04E-08
|
V
|
2.85E-02
|
0.02
|
0.98
|
0.03
|
3.74E-04
|
3.14E-08
|
Overall, upwelling of PBT organic pollutants can be modeled using the modified advective pumping theory and the two-stage desorption model. As expected, overlying stream water contaminant concentration increase rapidly when the desorption of contaminants from streambed sediments increase. Increase in the desorption of contaminants from the streambed, thereby, cause an increase in the uptake of PBT organic pollutants by benthic organisms. The overall findings of this project document an improved understanding and prediction capability for the transport and uptake of PBT organic pollutants present in sediments.
High flow condition. While keeping the same Rio Grande sediment in the streambed, two subsequent preliminary experimental runs (Preliminary Run #1 and Preliminary Run #2) were conducted for a time period of 45 hours. Both preliminary experiments were conducted in the same manner one following the other while leaving the sediment bed untouched. Thus, after Preliminary Run #1 was completed, the water was drained from the flume and clean DI water was reintroduced to the same height as that used in Preliminary Run #1 to begin Preliminary Run #2. The purpose for two consecutive runs with the same bed sediment was to examine if significant resuspension would still occur after Preliminary Run #1 (resuspension event). For each run, the experiment was conducted first using a low flow rate, which was then increased to a high flow rate after 40 minutes to examine the effects of stream flow on resuspension. Table 4 shows the hydraulic conditions used in each of the preliminary flume run.
Table 4. Hydraulic parameters used in each of the preliminary flume run.
Parameter
|
Low Flow Rate
|
High Flow Rate
|
Velocity (cm/s)
|
16.75
|
33.24
|
Underflow Velocity (cm/s)
|
1.31
|
2.59
|
Froude No.
|
0.20
|
0.40
|
E.G.L. Slope (cm/m)
|
0.04
|
0.17
|
Bed Friction Factor
|
0.07
|
0.07
|
Hydraulic Radius (m)
|
0.05
|
0.05
|
Bed Shear stress (N/m2)
|
0.21
|
0.82
|
In each run, the water column conductivity was measured and the total suspended solid concentration was only evaluated through UV-spectrophotometer absorbance measurements over time. For preliminary purposes, this was considered acceptable since increases in water conductivities indicate increases in the total dissolved solids in the suspended form. Additionally, increases in absorbance values indicate increases in the total suspended solids in the water column due to release from the streambed. Results from these preliminary experiments are shown in Figures 17 and 18.
As shown in Figure 17, the conductivity trend is similar as it increases over time in both preliminary runs. However, it should be noted that the hydraulic conductivity values for Preliminary Run #2 are significantly lower than those values for Preliminary Run #1. The most probable cause of this is that most of the dissolved solids had already been resuspended or released into the water column during Preliminary Run #1. The conductivity data suggests that there is continual release of total dissolved solids over time under both flow conditions for each run.
Figure 18 illustrates the results for the total suspended solids through UV-spectrophotometer absorbance values. The absorbance results indicate that there is an increase in the total suspended solids concentration in both experimental flume runs as the flow rate is increased (indicated by the red line). The absorbance values are much higher in Preliminary Run #1, which may be due to the same assumption that most of the fine sized sediments were resuspended and washed out during Preliminary Run #1. In addition, the absorbance results indicate that there is a significant increase in the total suspended solids concentrations at high flow rate.
Overall these preliminary results indicate that increases in flow rate do increase the total suspended solids in the water column over time. Additionally, bedform observations showed that there is pronounced bedform rearrangement within the first 40 mins of the experiment and as the flow rate is increased. This observation was very evident during Prelimanary Run #1 as seen in Figures 19A and 19B. Bedform changes within the first 40 minutes are considered as rearrangement of the bedforms due to the introduction of stream flow. After the 40 minutes, additional bedform rearrangement occurs due to the increased flow rate for Preliminary Run #1. On the other hand, no significant bedform movement was observed throughout the duration of the Preliminary Run #2 experiments as indicated by bedform measurements as seen in Figure 20.
One main benefit from these experiments is that the experimental protocol was tested and verified. Results from these experiments show that a new set of sediment should be used prior to every experiment since the sediment fines can be easily washed out. Additionally, sediment bed form stability can be achieved between flume runs if keeping the same sediment as shown from the bedform results. Furthermore, the total dissolved solids do not seem to be greatly influenced by increasing flow rates. However, increasing flow rates evidently increase the bed shear stress and cause an increase in the suspension of sediments as shown in Figure 18. Thus, the total suspension of solids is highly dependent on the stream flow rate.
References:
Ren J, Packman AI. Multi-phase contaminant exchange between streams and streambeds: theory and numerical simulations. Environmental Science & Technology 2004;38:2901-2911.Journal Articles:
No journal articles submitted with this report: View all 10 publications for this projectSupplemental Keywords:
sediment transport, bioassessment, sediment flux, PCB stream dynamics, sediment risk assessment.
, RFA, Scientific Discipline, Ecosystem Protection/Environmental Exposure & Risk, Aquatic Ecosystems & Estuarine Research, Aquatic Ecosystem, Environmental Monitoring, Ecological Risk Assessment, bioassessment, risk assessment, aquatic sediments, aquatic ecosystems, PCB fluxes, riverine ecosystems, sediment dynamics
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.